R/emmeans.R

In truthiness: Illusory Truth Longitudinal Study

#' Custom Contrast Function for emmeans
#'
#' Create a contrast matrix for equivalence test of a 2x4
#' interaction.
#'
#' @param levels Interaction levels (should be of length 8).
#'
#' @param ... Any other arguments (NB: currently ignored).
#'
#' @details  Runs all six ways of comparing the simple effects of
#' the two-level factor. For use with the emmeans package.
#'
#' @return A data frame to be passed to the \code{specs} argument of
#'   \code{\link[emmeans]{emmeans}}, each column of which represents
#'   predictor codings that contrast the illusory truth effect across
#'   two intervals.
#' 
#' @examples
#' library(ordinal)
#' library(emmeans)
#' 
#' ## create data frame with predictor codings
#' moddata <- get_model_data()
#' 
#' ## use 'allsimp' with emmeans for equivalence test
#' mod_emm <- emmeans(truth_trajectory_models[["ix2"]],
#'                    allsimp ~ Rep * Int, data = moddata)
#' 
#' mod_emm
#' 
#' @export
allsimp.emmc <- function(levels, ...) {
  if (length(levels) != 8L) stop("'levels' must have 8 levels")
  M <- data.frame("eff @ 1 day vs eff @ immediate" = c(1,-1,-1,1,0,0,0,0),
                  "eff @ 1 week vs eff @ immediate" = c(1, -1, 0, 0, -1, 1, 0, 0),
                  "eff @ 1 month vs eff @ immediate" = c(1, -1, 0, 0, 0, 0, -1, 1),
                  "eff @ 1 week vs eff @ 1 day" = c(0, 0, 1, -1, -1, 1, 0, 0),
                  "eff @ 1 month vs eff @ 1 day" = c(0, 0, 1, -1, 0, 0, -1, 1),
                  "eff @ 1 week vs eff @ 1 month" = c(0, 0, 0, 0, 1, -1, -1, 1),
                  check.names = FALSE)
  rownames(M) <- levels
  attr(M, "desc") <- "Simple effects comparison"
  M
}